The strategies above involve 5 steps that together comprise a push-pull approach, to optimize antigen structure, improve quantity and quality of the response, & remove regulatory barriers. Regulation of tumor immunity by NKT cells. NKT cells are true T cells restricted by a non-classical class I MHC molecule, CD1d, which presents glycolipid antigens. We discovered a novel immunoregulatory pathway in which NKT cells suppress tumor immunosurveillance, using IL-13 to induce myeloid cells to make TGF-b that suppresses immunity. Translating, we completed a phase I trial of anti-TGF-b as a new checkpoint blocker with clinical benefit in several melanoma patients. We discovered synergy between TGF-b blockade and 2 types of cancer vaccines in mice. We found that blockade of TGF-b1 and 2 (without 3) is sufficient to enhance immunosurveillance and vaccine efficacy and that this was further amplified by PD-1 blockade. However, type I (invariant TCR) NKT cells promote tumor immunity, whereas we found that type II (diverse TCR) NKT cells suppressed tumor immunity. These subsets cross-regulate each other, defining a new immunoregulatory axis. We are investigating the relationship between this novel NKT regulatory axis & other regulatory cells and molecules, including Treg cells, MDSCs and PD-1. Recently we found that both type II NKT cells and Treg cells can suppress tumor immunity concurrently, but type I inhibit type II NKT cells and leave Tregs as the dominant suppressor unless type I NKT cells are blocked or absent. Thus, the balance between 2 regulatory cells is determined by a 3rd cell that regulates the regulators. Moreover, we found different regulatory mechanisms in the same tumor in the lung and skin, even in the same mouse. The effector cells also cross between tissues in one direction only. Thus, tissue context determines cancer immunity even for the same tumor, implying that immunotherapy for primary tumors & metastases in different tissues may need to be different. We are performing structure-function studies of synthetic lipids to characterize ones that activate type I vs type II NKT cells. Further, we succeeded in making sulfatide-loaded CD1d tetramers that detect a subset of type II NKT cells in the liver & lung and have extensively characterized these cells, showing differences in surface markers, transcription factors & RNA expression profile. We are also testing the ability of sulfatide analogues to activate or inhibit type II NKT cell activities, both to understand function and to seek a specific antagonist to inhibit type II NKT cells. We identified a new class of agonist for type I NKT cells, b-ManCer, that inhibits tumors by a mechanism distinct from that of a-GalCer, requiring TNF-a and nitric oxide synthase instead of interferon (IFN)-g. This agonist also synergizes with a a-GalCer, is much less anergy-inducing than a-GalCer, and stimulates human NKT cells also. All of these studies are aimed to remove the negative regulatory roadblocks and/or improve the balance along the type I-II NKT axis to allow cancer vaccines to successfully induce tumor regression. Epitope enhancement and T-cell and antibody-based cancer vaccine strategies and translation to clinical trials. We carried out epitope enhancement (sequence modification to improve MHC binding) on an HLA-A2-binding epitope we discovered in a novel prostate and breast cancer antigen, TARP. The enhanced epitope induces human T cells that kill human tumor cells. We translated this to a phase I clinical trial of 2 peptides in stage D0 prostate cancer. 74% of vaccinees had a decreased PSA slope & tumor growth rate at 1 year (p = 0.0004). A randomized placebo-controlled phase II study is ongoing. We have studied a novel adenovirus-based HER-2 vaccine expressing the extracellular (EC) and transmembrane (TM) domains of rat neu (ErbB2), which prevents tumor growth in the neu-transgenic mice, and cures large established TUBO mammary tumors (2 cm) & large established lung metastases. The therapeutic effect in mice is purely antibody mediated, through inhibiting ErbB-2 function, unlike trastuzumab, which is FcR dependent, so may work in trastuzumab failures. We are carrying out a phase I/II trial of a cGMP human version of this vaccine, first examining safety & preliminary efficacy in HER2+ cancer in patients naive to trastuzumab, and then studying breast cancer patients with 3+ HER2 and prior trastuzumab. In Part 1, at the 2nd and 3rd dose levels, 5/11 (45%) evaluable patients showed clinical benefit, which, with safety, prompted the FDA to recently allow expansion to a 4th dose level and to HER2 3+ breast cancer patients who had progressed on HER2 therapy. We also carried out a CRADA-collaborative study in mice of an intratumoral therapy that we found induces a T cell response necessary for full regression and that results in long-term memory and resistance to rechallenge. Cytokines as vaccine adjuvants and induction of high avidity T cells. Our earlier work showed that high avidity T cells were more effective at clearing viral infections and cancers, and we found ways to induce them with cytokines and TLR ligands. The quality of response proved more important than the quantity. We recently found, using a novel adjuvant, CAF09, that we could lower the vaccine dose sufficiently to induce higher avidity CD4 T cells to more effectively clear virus infection. We also found that IL-1b can serve as a vaccine adjuvant but it induces Th17 helper cells that surprisingly do not work well to help Tc1 CD8 T cells that protect against vaccinia virus. Rather, they skew the CD8 response to Tc17 cells that make IL-17 and do not protect. Blockade of TGF-b can prevent this unwanted complication. We also discovered a role for IL-15 in skewing gut T helper subsets, through its effect on mononuclear phagocytes. Mucosal immunity, microbiome and HIV/SIV vaccines. About 85% of HIV transmission is mucosal. We found that a mucosal T cell vaccine can impact the initial mucosal nidus of infection. We are studying induction and trafficking of T cells, DCs, and MDSCs among mucosal compartments to optimize mucosal vaccine efficacy. In mice, we found that T cells could be directly primed in the vaginal mucosa, despite lack of organized lymphoid structures, contrary to textbook dogma. We also discovered that colonic DCs can program CD8 T cells to home back to the colon preferentially, based on differential retinoic acid expression vs. small intestine DCs, leading to imprinting of different receptors. We discovered that altering a cathepsin S cleavage site could protect an immunodominant epitope of gp120 from degradation in endosomes during cross-presentation, providing proof of concept for a novel mechanism of virus escape relevant for viruses like HIV that infect mostly non-APCs. Using NHP models, we found that activated mucosal T cells determine susceptibility to infection (transmission), eclipse time prior to systemic viral detection and acute viral load. We found that even in naive animals, gut microbiota can strongly affect susceptibility to transmission, through their effect on immune activation, and also affect vaccine efficacy. Further, we found that vaccines can induce MDSCs that counteract vaccine protection, and infection can affect trafficking of MDSCs. We also demonstrated for the first time that MDSCs could be infected by SHIV in vivo. We have translated our oral nanoparticle (NP) approach to macaque SIV vaccines and have found reduced risk against SHIV rectal acquisition in 2 studies. Mechanistic studies are ongoing. In addition, we are combining an SIV vaccine and mucosal NP boost to increase mucosal immunity with a microbicide to reduce the viral inoculum in an OAR-funded study.